A link to the full paper is here.

This paper is also published as a working paper at the Center for International Development at Harvard University.

Abstract: Nearly all challenges in international development tend to be complex because they depend on constantly evolving human behaviour, systems, and contexts, involving multiple actors, entities, and processes. As a result, both the discovery and scaling of innovations to address challenges in development often involve changes in system behaviour or even system-level transformation.  This is rarely a linear process over time and can result in unexpected outcomes. Existing evaluation techniques commonly used in international development, including Randomized Control Trials (RCT) and quasi-experimental methods, are good at assessing specific effects of interventions but are not designed for the change processes inherent to innovation and scaling within a system. There is a need to reconstruct how we use existing measurement tools, techniques, and methodologies so that they capture the complexity of the environment in which an intervention or change occurs.

We introduce Adaptive Evaluation, designed to learn at various levels of complexity while supporting the transformation needed to foster sustainable change. An Adaptive Evaluation uses three main approaches to work with complex questions—systems diagnosis, theory-based assessment of change processes, and iterative designs. An Adaptive Evaluation typically builds hypotheses from field-based interactions, emphasizes learning over testing, advocates open-mindedness with techniques, and appreciates the value of dialogue and participation in navigating complex processes. It can use RCT or similar techniques to analyse specific processes within a system or a development cycle, but these are embedded in a broader approach to assessment and interpretation.  It is designed to be flexible and adjust to shifting contexts. Finally, an Adaptive Evaluation can be applied at any stage in a complex intervention’s lifecycle, from the interpretation of the system and change processes to rapid experimentation, prototyping, and testing of select interventions, and then adaptation to different settings for impact at scale. This paper provides the theoretical basis for an Adaptive Evaluation—the main approaches, core ideology, process, and applications.